What Is the Matter with Grading in the Age of AI?
Source: Nick Potkalitsky Substack
Author: Nick Potkalitsky
Original source: https://nickpotkalitsky.substack.com/p/what-is-the-matter-with-grading-in
Published: 2026-03-22
Source type: essay
Private backup: the full article text is archived in the private repository at archives/articles/nickpotkalitsky-substack-com-what-is-the-matter-with-grading-in.source.md. It is not published on the public Quartz site.
Summary
Potkalitsky recounts working with an eighth-grade student whose AI-powered writing feedback tool gave her original formulaic essay a 100%, then gave a stronger revised version an 80%, and then a 90% when rerun in a new chat. The essay had improved through genuine human instruction, but the inconsistent AI scoring undermined the student’s confidence in revision, the teacher’s approach, the state-testing rubric, and her own judgment as a writer. The article argues that unreliable AI evaluation can destabilize the feedback ecosystem around learning, encouraging students to optimize for the machine’s most forgiving score rather than trust intellectual revision.
Pull quotes
Authority without precision
“The tool presents with the same confidence regardless of its actual precision.”
Destabilizing feedback
“It actively destabilizes the feedback ecosystem that learning depends on.”
Compliance after failure
“She is twelve years old, and she has already learned that lesson.”
No reliable referee
“School becomes a game you are winning or losing, and the game has no reliable referee.”
Durability note
The specific scoring example is tied to one tool encounter, but the durable lesson is broader: AI feedback systems can change how students interpret authority, revision, and trust in assessment.
Big ideas
- AI tools should be judged by the work they will actually do
- Schools should start with learning values before choosing AI tools
- Learning still needs some struggle, even when AI can make things easier
- AI literacy should help people notice how AI changes what counts as knowing
- Students need to bring the purpose; AI should not supply it for them
Claims
- AI grading and feedback systems need validation before schools trust them
- AI grading systems need transparency, validation, and bias checks
- In an AI world, assessment should focus on watching students think
- AI can make school feel more transactional
Key evidence and examples
- A formulaic essay scored 100% by an AI feedback tool.
- Human tutoring improved the counterargument and statistical reasoning, but the revised essay scored 80% and then 90% on rerun.
- The AI-generated feedback was vague and misquoted the student’s actual essay.
- The student restored the original 100% version and submitted it, illustrating erosion of trust in effort, quality, and outcome.
Education relevance
Very high for AI grading, writing instruction, classroom feedback, student motivation, assessment design, educational trust, and automated evaluation policy.